Predictors of Acute Myocardial Infarction: A Machine Learning Analysis After a 7-Year Follow-Up

Background: Ischemic heart disease is a major global health problem with significant morbidity and mortality. Several cardiometabolic variables play a key role in the incidence of adverse cardiovascular outcomes. Objectives: The aim of the present study was to apply a machine learning approach to investigate factors that can predict acute coronary syndrome in patients with a previous episode. Methods: We recruited 652 patients, admitted to the hospital for acute coronary syndrome, eligible if undergoing immediate coronary revascularization procedures for ST-segment-elevation myocardial infarction or coronary revascularization procedures within 24 h. Results: Baseline pulse wave velocity appears to be the most predictive variable overall, followed by the occurrence of left ventricular hypertrophy and left ventricular end-diastolic diameters. We found that the potential of machine learning to predict life-threatening events is significant. Conclusions: Machine learning algorithms can be used to create models to identify patients at risk for acute myocardial infarction. However, great care must be taken with data quality and ethical use of these algorithms.